Multi-touch attribution tools: what they can and cannot tell you
For the full comparison, see MMM vs MTA vs incrementality tests. We do not rank or rate vendors.
What do multi-touch attribution tools do?
Multi-touch attribution (MTA) assigns credit to several touchpoints on the path to a purchase. A tool collects those touchpoints, links them to orders, and applies a rule.
Attribution models are those rules. Last-click attribution gives everything to the final click. Data-driven attribution is Google's model, built from converting and non-converting paths. Other tools offer their own versions.
One point matters more than the rest. Attribution is a rule for sharing credit. It is not proof of cause. One agency's measurement guide puts it plainly: attribution gives speed, an incrementality test gives proof of cause, and marketing-mix modelling gives the cross-channel view1. It also says attribution cannot tell you whether the conversion would have happened without the ad1.
Why do e-commerce brands use them?
Example (illustrative): you run Meta, Google and email. Each platform claims the same order. You want one view that does not add up to 140% of your revenue.
That is the pull. A tool gives you one place, quicker than a test and more detailed than a total. It can show which ads and campaigns appear on converting paths.
Why do the numbers differ in the first place? A guide from a vendor that sells an attribution tool says each source counts something different: the shop counts orders, ad platforms count touches inside their own window, GA4 counts the sessions it managed to follow2. An attribution tool picks one rule and applies it to everything. That makes the numbers consistent. It does not make them true.
Is there a best attribution tool for ecommerce and DTC brands?
Lists of "best attribution tools" are often written by companies that sell one. One such list, published by an attribution vendor, puts its own product first4. We do not rank tools here, and we do not rate vendors.
Compare tools on criteria instead:
- Which attribution rule can you see and change?
- Does it pull in order data from your shop, or only platform data?
- Can you export raw data to your own data warehouse?
- Does it help you run a test, or only report?
- Who owns the account and the data if you leave?
How do we implement it?
Check tracking first
Make sure tracking is working and not duplicated. The tool can only credit what it sees.
Choose the rule
Pick a tool where you can see and change the rule, and take the raw data to your own warehouse. Accounts sit with you.
Name a reader
Someone owns the reading, including why it disagrees with each platform.
Keep it a daily read
Use it to watch, not to decide. Moving budget between channels needs a test or a calibrated model first.
When should you not use them?
Skip them, for now, when:
If you want a second opinion first, you can start with a free measurement audit. Our article on Meta ROAS vs Shopify revenue covers the usual disagreement.
What does it cost in effort?
We give no figures.
The effort is mostly:
- Tracking quality. The tool can only credit what it sees. A path it cannot see is missing from the model.
- Set-up. Connecting shop, ad platforms and email, and agreeing the rule.
- Reading it. Someone must explain why it disagrees with each platform.
- Access. Our view: accounts and data should sit with you.
Pitfalls
Treating credit as cause
"Attribution shows the ads caused it" is not what the method says.
Another number nobody reconciles
A fourth source with no owner adds confusion.
Budget moves without a test
Credit shares are not a measure of what each channel added.
Coverage gaps
Consent choices and browser limits can hide part of the path. Your counsel decides what you may collect.
Our take: Watching
Our view: Watching. This is our rating, not a market fact or an industry scale.
The reason is what the method cannot see. Attribution is observational: it looks at paths and does not run an experiment. A 2019 paper compared 15 US advertising experiments at Facebook with observational methods. The observational methods often failed to produce the same effects as the experiments3. That is evidence about those experiments, published in 2019. It is not a measurement of today's Meta, of Europe, or of any specific attribution tool.
Our view: a tool is fine for a daily read. Moving budget between channels on its numbers alone needs a test or a calibrated model first.
FAQ
Is multi-touch attribution the same as incrementality?
No. MTA splits credit among touchpoints. Incrementality asks what would have happened without the ads.
How is a third-party attribution tool different from GA4?
One vendor's guide says GA4 counts only the sessions it can follow, while the vendor's own tool joins each shop order with platform touches and splits credit once2. Different inputs and rules, different numbers.
How do I choose between two attribution tools, such as Northbeam and Triple Whale?
Use the criteria above, on your own data. We do not rate vendors, and a comparison written by one of the two is sales material.
Can I trust the numbers from an attribution tool?
Trust them for direction and consistency. For budget moves, check them against a test.
Do I need an attribution tool at all?
Not automatically. Our view: many budget decisions can be made with platform reports, shop revenue and the occasional test.
Sources
- A Guide to Measurement in 2026 (Silverback Strategies)Vendor
- Why Meta, Google, GA4 and Shopify Never Match (TrueROAS)Vendor
- Gordon, Zettelmeyer, Bhargava, Chapsky: A Comparison of Approaches to Advertising Measurement (Marketing Science 2019) - strona KelloggPrimary source
- Best attribution tools for ecommerce (Cometly)Vendor
Let’s find what’s actually working.
Next step. Want a view on whether a tool belongs on your account? Talk to us about your budget decision. Or read about our Measurement & incrementality service.